Every complex equation is a set of two real equations: the one with the real parts and the other with imaginary parts:
$A = B \iff \begin{cases}
\operatorname{Re}(A) = \operatorname{Re}(B)\\
\operatorname{Im}(A) = \operatorname{Im}(B)\\
\end{cases}$
As long as we add complex equations together and/or multiply both sides by a real number, the real and imaginary parts may be treated strictly apart from each other; they do not mix.
The real current may be defined as a real part of some "complex current" $I$ and all those calculations will be valid regardless of what the imaginary part is. This is the boring answer to "how does it work?" question. More interesting is: why do we do that? As for now, the imaginary part seems to be unnecessary burden.
The trick is we consider the current in a form $I= I_0 e^{i \omega t + \varphi}$.
With that representation we have:
$\frac{d}{dt} I = i \omega I$
$I = \frac{d}{dt} \frac {-i} \omega I$
It means we can (for fixed $\omega$) replace the derivative with simple multiplication by a factor, and the integral (i.e. electric charge) with multiplication by another factor.
Of course multiplying by $i$ mixes real and imaginary parts, but that's the point. The imaginary part is not a burden anymore; it is derivative and integral (with respective factors) in one.
The formula for $Y$ in your example uses that with complex $V$. The relation between (real) $I$ and (real) $V$ in RLC circuit gives 2nd order differential equation. With complex numbers and fixed $\omega$ we can get non-differential equation:
$I = YV$
The imaginary part of $Y$ is a clever substitute for differentiation and integration.